summaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
Diffstat
-rw-r--r--Lib/test/test_codeccallbacks.py4
-rw-r--r--Lib/test/test_codecs.py84
-rw-r--r--Misc/NEWS2
-rw-r--r--Objects/unicodeobject.c82
4 files changed, 118 insertions, 54 deletions
diff --git a/Lib/test/test_codeccallbacks.py b/Lib/test/test_codeccallbacks.py
index 81bf80d..fd88505 100644
--- a/Lib/test/test_codeccallbacks.py
+++ b/Lib/test/test_codeccallbacks.py
@@ -271,12 +271,12 @@ class CodecCallbackTest(unittest.TestCase):
self.assertEqual(
b"\\u3042\u3xxx".decode("unicode-escape", "test.handler1"),
- "\u3042[<92><117><51><120>]xx"
+ "\u3042[<92><117><51>]xxx"
)
self.assertEqual(
b"\\u3042\u3xx".decode("unicode-escape", "test.handler1"),
- "\u3042[<92><117><51><120><120>]"
+ "\u3042[<92><117><51>]xx"
)
self.assertEqual(
diff --git a/Lib/test/test_codecs.py b/Lib/test/test_codecs.py
index 67690b8..4c0c6de 100644
--- a/Lib/test/test_codecs.py
+++ b/Lib/test/test_codecs.py
@@ -21,6 +21,11 @@ except ImportError:
else:
SIZEOF_WCHAR_T = ctypes.sizeof(ctypes.c_wchar)
+def coding_checker(self, coder):
+ def check(input, expect):
+ self.assertEqual(coder(input), (expect, len(input)))
+ return check
+
class Queue(object):
"""
queue: write bytes at one end, read bytes from the other end
@@ -2009,6 +2014,85 @@ class TypesTest(unittest.TestCase):
self.assertRaises(UnicodeDecodeError, codecs.raw_unicode_escape_decode, br"\U00110000")
self.assertEqual(codecs.raw_unicode_escape_decode(r"\U00110000", "replace"), ("\ufffd", 10))
+
+class UnicodeEscapeTest(unittest.TestCase):
+ def test_empty(self):
+ self.assertEqual(codecs.unicode_escape_encode(""), (b"", 0))
+ self.assertEqual(codecs.unicode_escape_decode(b""), ("", 0))
+
+ def test_raw_encode(self):
+ encode = codecs.unicode_escape_encode
+ for b in range(32, 127):
+ if b != b'\\'[0]:
+ self.assertEqual(encode(chr(b)), (bytes([b]), 1))
+
+ def test_raw_decode(self):
+ decode = codecs.unicode_escape_decode
+ for b in range(256):
+ if b != b'\\'[0]:
+ self.assertEqual(decode(bytes([b]) + b'0'), (chr(b) + '0', 2))
+
+ def test_escape_encode(self):
+ encode = codecs.unicode_escape_encode
+ check = coding_checker(self, encode)
+ check('\t', br'\t')
+ check('\n', br'\n')
+ check('\r', br'\r')
+ check('\\', br'\\')
+ for b in range(32):
+ if chr(b) not in '\t\n\r':
+ check(chr(b), ('\\x%02x' % b).encode())
+ for b in range(127, 256):
+ check(chr(b), ('\\x%02x' % b).encode())
+ check('\u20ac', br'\u20ac')
+ check('\U0001d120', br'\U0001d120')
+
+ def test_escape_decode(self):
+ decode = codecs.unicode_escape_decode
+ check = coding_checker(self, decode)
+ check(b"[\\\n]", "[]")
+ check(br'[\"]', '["]')
+ check(br"[\']", "[']")
+ check(br"[\\]", r"[\]")
+ check(br"[\a]", "[\x07]")
+ check(br"[\b]", "[\x08]")
+ check(br"[\t]", "[\x09]")
+ check(br"[\n]", "[\x0a]")
+ check(br"[\v]", "[\x0b]")
+ check(br"[\f]", "[\x0c]")
+ check(br"[\r]", "[\x0d]")
+ check(br"[\7]", "[\x07]")
+ check(br"[\8]", r"[\8]")
+ check(br"[\78]", "[\x078]")
+ check(br"[\41]", "[!]")
+ check(br"[\418]", "[!8]")
+ check(br"[\101]", "[A]")
+ check(br"[\1010]", "[A0]")
+ check(br"[\x41]", "[A]")
+ check(br"[\x410]", "[A0]")
+ check(br"\u20ac", "\u20ac")
+ check(br"\U0001d120", "\U0001d120")
+ for b in range(256):
+ if b not in b'\n"\'\\abtnvfr01234567xuUN':
+ check(b'\\' + bytes([b]), '\\' + chr(b))
+
+ def test_decode_errors(self):
+ decode = codecs.unicode_escape_decode
+ for c, d in (b'x', 2), (b'u', 4), (b'U', 4):
+ for i in range(d):
+ self.assertRaises(UnicodeDecodeError, decode,
+ b"\\" + c + b"0"*i)
+ self.assertRaises(UnicodeDecodeError, decode,
+ b"[\\" + c + b"0"*i + b"]")
+ data = b"[\\" + c + b"0"*i + b"]\\" + c + b"0"*i
+ self.assertEqual(decode(data, "ignore"), ("[]", len(data)))
+ self.assertEqual(decode(data, "replace"),
+ ("[\ufffd]\ufffd", len(data)))
+ self.assertRaises(UnicodeDecodeError, decode, br"\U00110000")
+ self.assertEqual(decode(br"\U00110000", "ignore"), ("", 10))
+ self.assertEqual(decode(br"\U00110000", "replace"), ("\ufffd", 10))
+
+
class SurrogateEscapeTest(unittest.TestCase):
def test_utf8(self):
diff --git a/Misc/NEWS b/Misc/NEWS
index b8de2f0..f886005 100644
--- a/Misc/NEWS
+++ b/Misc/NEWS
@@ -162,6 +162,8 @@ Core and Builtins
Library
-------
+- Issue #16979: Fix error handling bugs in the unicode-escape-decode decoder.
+
- Issue #1602133: on Mac OS X a shared library build (``--enable-shared``)
now fills the ``os.environ`` variable correctly.
diff --git a/Objects/unicodeobject.c b/Objects/unicodeobject.c
index c96a91c..b559cb1 100644
--- a/Objects/unicodeobject.c
+++ b/Objects/unicodeobject.c
@@ -5508,7 +5508,6 @@ PyUnicode_DecodeUnicodeEscape(const char *s,
const char *starts = s;
Py_ssize_t startinpos;
Py_ssize_t endinpos;
- int j;
PyObject *v;
const char *end;
char* message;
@@ -5630,29 +5629,19 @@ PyUnicode_DecodeUnicodeEscape(const char *s,
message = "truncated \\UXXXXXXXX escape";
hexescape:
chr = 0;
- if (s+digits>end) {
- endinpos = size;
- if (unicode_decode_call_errorhandler(
- errors, &errorHandler,
- "unicodeescape", "end of string in escape sequence",
- &starts, &end, &startinpos, &endinpos, &exc, &s,
- &v, &i))
- goto onError;
- goto nextByte;
- }
- for (j = 0; j < digits; ++j) {
- c = (unsigned char) s[j];
- if (!Py_ISXDIGIT(c)) {
- endinpos = (s+j+1)-starts;
- if (unicode_decode_call_errorhandler(
- errors, &errorHandler,
- "unicodeescape", message,
- &starts, &end, &startinpos, &endinpos, &exc, &s,
- &v, &i))
- goto onError;
- len = PyUnicode_GET_LENGTH(v);
- goto nextByte;
+ if (end - s < digits) {
+ /* count only hex digits */
+ for (; s < end; ++s) {
+ c = (unsigned char)*s;
+ if (!Py_ISXDIGIT(c))
+ goto error;
}
+ goto error;
+ }
+ for (; digits--; ++s) {
+ c = (unsigned char)*s;
+ if (!Py_ISXDIGIT(c))
+ goto error;
chr = (chr<<4) & ~0xF;
if (c >= '0' && c <= '9')
chr += c - '0';
@@ -5661,24 +5650,16 @@ PyUnicode_DecodeUnicodeEscape(const char *s,
else
chr += 10 + c - 'A';
}
- s += j;
if (chr == 0xffffffff && PyErr_Occurred())
/* _decoding_error will have already written into the
target buffer. */
break;
store:
/* when we get here, chr is a 32-bit unicode character */
- if (chr <= MAX_UNICODE) {
- WRITECHAR(chr);
- } else {
- endinpos = s-starts;
- if (unicode_decode_call_errorhandler(
- errors, &errorHandler,
- "unicodeescape", "illegal Unicode character",
- &starts, &end, &startinpos, &endinpos, &exc, &s,
- &v, &i))
- goto onError;
- }
+ message = "illegal Unicode character";
+ if (chr > MAX_UNICODE)
+ goto error;
+ WRITECHAR(chr);
break;
/* \N{name} */
@@ -5706,26 +5687,13 @@ PyUnicode_DecodeUnicodeEscape(const char *s,
goto store;
}
}
- endinpos = s-starts;
- if (unicode_decode_call_errorhandler(
- errors, &errorHandler,
- "unicodeescape", message,
- &starts, &end, &startinpos, &endinpos, &exc, &s,
- &v, &i))
- goto onError;
- break;
+ goto error;
default:
if (s > end) {
message = "\\ at end of string";
s--;
- endinpos = s-starts;
- if (unicode_decode_call_errorhandler(
- errors, &errorHandler,
- "unicodeescape", message,
- &starts, &end, &startinpos, &endinpos, &exc, &s,
- &v, &i))
- goto onError;
+ goto error;
}
else {
WRITECHAR('\\');
@@ -5733,8 +5701,18 @@ PyUnicode_DecodeUnicodeEscape(const char *s,
}
break;
}
- nextByte:
- ;
+ continue;
+
+ error:
+ endinpos = s-starts;
+ if (unicode_decode_call_errorhandler(
+ errors, &errorHandler,
+ "unicodeescape", message,
+ &starts, &end, &startinpos, &endinpos, &exc, &s,
+ &v, &i))
+ goto onError;
+ len = PyUnicode_GET_LENGTH(v);
+ continue;
}
#undef WRITECHAR
generated by cgit v0.12 at 2024-12-03 21:23:32 (GMT)